The long-term goal of this research project component is to develop effective methods of preventing and reducing the problem of relapse in adolescents who have modified their substance abuse. The project has four specific aims: (1) To describe long-term *up to 5 years) alcohol (and other drug) use and other psychosocial functioning (e.g., education, use of free, time, peer relationships) areas in the current and new PAARC clinical samples; (2) To evaluate the adequacy of a variant of a proposed model of outcome following treatment adapted from the adult literature for the explanation of substance use and long-term functioning in other areas, and of empirically derived definitions of relapse in the PAARC clinical samples; (3) To study in the new clinical sample the immediate precipitant of alcohol or other drug use lapses/relapses, in the context of person, treatment, and environmental determinants; and (4) To conduct event history analyses of lapses and relapses in the new clinical sample. Subjects will be males and females from the current and new PAARC clinical samples who at baseline meet criteria for DSM-IV current alcohol abuse or dependence diagnosis. It is estimated that 162, 150, and 150 adolescent in equal numbers of males and females will have met criteria for inclusion in this project from the current PAARC cohort and will have completed 1-,3-, and 5-year follow-up assessments, respectively. Counterpart numbers for the new clinical cohort are 130 and 60 for the 1- and 3-year follow-ups. Where relevant, gender- and age-matched community controls also will be sampled from the clinical core for purposes of analysis. In addition to the standard PAARC assessment protocol, new cohort subjects in this component will complete additional measures required to achieve this project's specific aim. In addition, new cohort clinical subjects in this study will be administered monthly telephone interviews for the first 12 months post baseline that focus on daily alcohol and other drug use, treatment activities, and relapse precipitants. Data analyses will include descriptive statistics, comparison of group means by use of univariate and multivariate tests, path analysis, and event history analyses.